Cheshire Cat AI API Client
API Client made in TypeScript to communicate with the Cheshire Cat AI.
The package provides a class to interface with the Cheshire Cat Python Server.
It can be used both in Browser or NodeJS environment.
Every endpoint is a CancelablePromise
, it means you can cancel the request if you want.
Installation
npm install ccat-api
yarn add ccat-api
pnpm i ccat-api
Getting started
To set up the client, you must first import the CatClient
class:
import { CatClient } from 'ccat-api'
const cat = new CatClient({
baseUrl: 'localhost',
})
Client settings
API_KEY, CORE_HOST, CORE_PORT and CORE_USE_SECURE_PROTOCOLS refer to the CCAT Core .env file.
Property | Type | Default | Description |
---|
baseUrl | string | Required | The same of CORE_HOST |
authKey | string | '' | The same of API_KEY |
port | string | '1865' | The same of the CORE_PORT |
instant | boolean | true | Instantly initialize the websocket and the API client, or later with .init() |
secure | boolean | false | The same of the CORE_USE_SECURE_PROTOCOLS |
timeout | number | 10000 | Timeout for the endpoints, in milliseconds |
ws | string | undefined | An object of type WebSocketSettings |
WebSocket settings
Property | Type | Default | Description |
---|
path | string | 'ws' | Websocket path to use to communicate with the CCat |
retries | number | 3 | The maximum number of retries before calling onFailed |
delay | number | 5000 | The delay for reconnect, in milliseconds |
onFailed | (ErrorCode) => void | undefined | The function to call after failing all the retries |
Then, for example, you can configure the LLM like this:
cat.api.settingsLargeLanguageModel.upsertLlmSetting('LLMOpenAIConfig', {
openai_api_key: 'OPEN_API_KEY'
})
To send a message to the cat, you can:
cat.send('Hello my friend!')
You can listen to the websocket events:
cat.onConnected(() => {
console.log('Socket connected')
}).onMessage(msg => {
console.log(msg)
}).onError(err => {
console.log(err)
}).onDisconnected(() => {
console.log('Socket disconnected')
})
For example, you can get the list of plugins like this:
cat.api.plugins.listAvailablePlugins().then(plugins => {
console.log(plugins)
})
Credits
Made for the Cheshire Cat AI organization.
This was possible thanks to openapi-typescript-codegen.